zqz/platform-external-webrtc
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

AGC2 lightweight noise floor estimator

Browse Source 
The current noise level estimator has a bug due to which the estimated
level decays to the lower bound in a few seconds when speech is observed.
Instead of fixing the current implementation, which is based on a
stationarity classifier, an alternative, lightweight, noise floor
estimator has been added and tuned for AGC2.

Tested on several AEC dumps including HW mute, music and fast talking.

Bug: webrtc:7494
Change-Id: Iae4cff9fc955a716878f830957e893cd5bc59446
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/214133
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Reviewed-by: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33733}
This commit is contained in:
Alessio Bazzica
2021-04-14 16:17:09 +02:00
committed by Commit Bot
parent 3ab7a55f6e
commit 61982a7f2d
10 changed files with 246 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
modules/audio_processing/include/audio_processing.h
View File

@ -350,21 +350,23 @@ class RTC_EXPORT AudioProcessing : public rtc::RefCountInterface {
}
enum LevelEstimator { kRms, kPeak };
enum NoiseEstimator { kStationaryNoise, kNoiseFloor };
bool enabled = false;
struct FixedDigital {
float gain_db = 0.0f;
} fixed_digital;
struct AdaptiveDigital {
bool enabled = false;
NoiseEstimator noise_estimator = kNoiseFloor;
int vad_reset_period_ms = 1500;
float vad_probability_attack = 0.3f;
float vad_probability_attack = 0.9f;
LevelEstimator level_estimator = kRms;
int level_estimator_adjacent_speech_frames_threshold = 6;
int level_estimator_adjacent_speech_frames_threshold = 11;
// TODO(crbug.com/webrtc/7494): Remove `use_saturation_protector`.
bool use_saturation_protector = true;
float initial_saturation_margin_db = 20.0f;
float extra_saturation_margin_db = 5.0f;
int gain_applier_adjacent_speech_frames_threshold = 6;
int gain_applier_adjacent_speech_frames_threshold = 11;
float max_gain_change_db_per_second = 3.0f;
float max_output_noise_level_dbfs = -55.0f;
bool sse2_allowed = true;